Hash cleanup.

This commit is contained in:
Бранимир Караџић
2022-05-23 18:45:58 -07:00
parent c634760181
commit 0314d3b018
4 changed files with 261 additions and 198 deletions

View File

@@ -11,45 +11,7 @@
namespace bx
{
/// MurmurHash2 was written by Austin Appleby, and is placed in the public
/// domain. The author hereby disclaims copyright to this source code.
///
class HashMurmur2A
{
public:
///
void begin(uint32_t _seed = 0);
///
void add(const void* _data, int _len);
///
template<typename Ty>
void add(Ty _value);
///
uint32_t end();
private:
///
void addAligned(const void* _data, int _len);
///
void addUnaligned(const void* _data, int _len);
///
static void readUnaligned(const void* _data, uint32_t& _out);
///
void mixTail(const uint8_t*& _data, int& _len);
uint32_t m_hash;
uint32_t m_tail;
uint32_t m_count;
uint32_t m_size;
};
///
/// 32-bit Adler checksum hash.
class HashAdler32
{
public:
@@ -57,11 +19,17 @@ namespace bx
void begin();
///
void add(const void* _data, int _len);
void add(const void* _data, int32_t _len);
///
void add(const char* _data);
///
void add(const StringView& _data);
///
template<typename Ty>
void add(Ty _value);
void add(const Ty& _data);
///
uint32_t end();
@@ -71,7 +39,7 @@ namespace bx
uint32_t m_b;
};
///
/// 32-bit cyclic redundancy checksum hash.
class HashCrc32
{
public:
@@ -88,11 +56,17 @@ namespace bx
void begin(Enum _type = Ieee);
///
void add(const void* _data, int _len);
void add(const void* _data, int32_t _len);
///
void add(const char* _data);
///
void add(const StringView& _data);
///
template<typename Ty>
void add(Ty _value);
void add(const Ty& _data);
///
uint32_t end();
@@ -102,21 +76,51 @@ namespace bx
uint32_t m_hash;
};
/// 32-bit multiply and rotate hash.
class HashMurmur2A
{
public:
///
void begin(uint32_t _seed = 0);
///
void add(const void* _data, int32_t _len);
///
void add(const char* _data);
///
void add(const StringView& _data);
///
template<typename Ty>
void add(const Ty& _data);
///
uint32_t end();
private:
uint32_t m_hash;
uint32_t m_tail;
uint32_t m_count;
uint32_t m_size;
};
///
template<typename HashT>
uint32_t hash(const void* _data, uint32_t _size);
///
template<typename HashT, typename Ty>
uint32_t hash(const Ty& _data);
template<typename HashT>
uint32_t hash(const char* _data);
///
template<typename HashT>
uint32_t hash(const StringView& _data);
///
template<typename HashT>
uint32_t hash(const char* _data);
template<typename HashT, typename Ty>
uint32_t hash(const Ty& _data);
} // namespace bx

View File

@@ -9,142 +9,16 @@
namespace bx
{
#define MURMUR_M 0x5bd1e995
#define MURMUR_R 24
#define mmix(_h, _k) { _k *= MURMUR_M; _k ^= _k >> MURMUR_R; _k *= MURMUR_M; _h *= MURMUR_M; _h ^= _k; }
inline void HashMurmur2A::begin(uint32_t _seed)
{
m_hash = _seed;
m_tail = 0;
m_count = 0;
m_size = 0;
}
inline void HashMurmur2A::add(const void* _data, int _len)
{
if (BX_UNLIKELY(!isAligned(_data, 4) ) )
{
addUnaligned(_data, _len);
return;
}
addAligned(_data, _len);
}
inline void HashMurmur2A::addAligned(const void* _data, int _len)
{
const uint8_t* data = (const uint8_t*)_data;
m_size += _len;
mixTail(data, _len);
while(_len >= 4)
{
uint32_t kk = *(uint32_t*)data;
mmix(m_hash, kk);
data += 4;
_len -= 4;
}
mixTail(data, _len);
}
inline void HashMurmur2A::addUnaligned(const void* _data, int _len)
{
const uint8_t* data = (const uint8_t*)_data;
m_size += _len;
mixTail(data, _len);
while(_len >= 4)
{
uint32_t kk;
readUnaligned(data, kk);
mmix(m_hash, kk);
data += 4;
_len -= 4;
}
mixTail(data, _len);
}
template<typename Ty>
inline void HashMurmur2A::add(Ty _value)
{
add(&_value, sizeof(Ty) );
}
inline uint32_t HashMurmur2A::end()
{
mmix(m_hash, m_tail);
mmix(m_hash, m_size);
m_hash ^= m_hash >> 13;
m_hash *= MURMUR_M;
m_hash ^= m_hash >> 15;
return m_hash;
}
inline void HashMurmur2A::readUnaligned(const void* _data, uint32_t& _out)
{
const uint8_t* data = (const uint8_t*)_data;
if (BX_ENABLED(BX_CPU_ENDIAN_BIG) )
{
_out = 0
| data[0]<<24
| data[1]<<16
| data[2]<<8
| data[3]
;
}
else
{
_out = 0
| data[0]
| data[1]<<8
| data[2]<<16
| data[3]<<24
;
}
}
inline void HashMurmur2A::mixTail(const uint8_t*& _data, int& _len)
{
while( _len && ((_len<4) || m_count) )
{
m_tail |= (*_data++) << (m_count * 8);
m_count++;
_len--;
if(m_count == 4)
{
mmix(m_hash, m_tail);
m_tail = 0;
m_count = 0;
}
}
}
#undef MURMUR_M
#undef MURMUR_R
#undef mmix
inline void HashAdler32::begin()
{
m_a = 1;
m_b = 0;
}
inline void HashAdler32::add(const void* _data, int _len)
inline void HashAdler32::add(const void* _data, int32_t _len)
{
const uint32_t kModAdler = 65521;
constexpr uint32_t kModAdler = 65521;
const uint8_t* data = (const uint8_t*)_data;
for (; _len != 0; --_len)
{
@@ -153,10 +27,20 @@ namespace bx
}
}
template<typename Ty>
inline void HashAdler32::add(Ty _value)
inline void HashAdler32::add(const char* _data)
{
add(&_value, sizeof(Ty) );
return add(StringView(_data) );
}
inline void HashAdler32::add(const StringView& _data)
{
return add(_data.getPtr(), _data.getLength() );
}
template<typename Ty>
inline void HashAdler32::add(const Ty& _data)
{
add(&_data, sizeof(Ty) );
}
inline uint32_t HashAdler32::end()
@@ -164,10 +48,20 @@ namespace bx
return m_a | (m_b<<16);
}
template<typename Ty>
inline void HashCrc32::add(Ty _value)
inline void HashCrc32::add(const char* _data)
{
add(&_value, sizeof(Ty) );
return add(StringView(_data) );
}
inline void HashCrc32::add(const StringView& _data)
{
return add(_data.getPtr(), _data.getLength() );
}
template<typename Ty>
inline void HashCrc32::add(const Ty& _data)
{
add(&_data, sizeof(Ty) );
}
inline uint32_t HashCrc32::end()
@@ -176,15 +70,51 @@ namespace bx
return m_hash;
}
inline void HashMurmur2A::begin(uint32_t _seed)
{
m_hash = _seed;
m_tail = 0;
m_count = 0;
m_size = 0;
}
inline void HashMurmur2A::add(const char* _data)
{
return add(StringView(_data) );
}
inline void HashMurmur2A::add(const StringView& _data)
{
return add(_data.getPtr(), _data.getLength() );
}
template<typename Ty>
inline void HashMurmur2A::add(const Ty& _data)
{
add(&_data, sizeof(Ty) );
}
template<typename HashT>
inline uint32_t hash(const void* _data, uint32_t _size)
{
HashT hh;
hh.begin();
hh.add(_data, (int)_size);
hh.add(_data, (int32_t)_size);
return hh.end();
}
template<typename HashT>
inline uint32_t hash(const char* _data)
{
return hash<HashT>(StringView(_data) );
}
template<typename HashT>
inline uint32_t hash(const StringView& _data)
{
return hash<HashT>(_data.getPtr(), _data.getLength() );
}
template<typename HashT, typename Ty>
inline uint32_t hash(const Ty& _data)
{
@@ -192,16 +122,4 @@ namespace bx
return hash<HashT>(&_data, sizeof(Ty) );
}
template<typename HashT>
inline uint32_t hash(const StringView& _data)
{
return hash<HashT>(_data.getPtr(), _data.getLength() );
}
template<typename HashT>
inline uint32_t hash(const char* _data)
{
return hash<HashT>(StringView(_data) );
}
} // namespace bx

View File

@@ -133,7 +133,7 @@ void HashCrc32::begin(Enum _type)
m_table = s_crcTable[_type];
}
void HashCrc32::add(const void* _data, int _len)
void HashCrc32::add(const void* _data, int32_t _len)
{
const uint8_t* data = (const uint8_t*)_data;
@@ -147,4 +147,136 @@ void HashCrc32::add(const void* _data, int _len)
m_hash = hash;
}
struct HashMurmur2APod
{
uint32_t m_hash;
uint32_t m_tail;
uint32_t m_count;
uint32_t m_size;
};
BX_STATIC_ASSERT(sizeof(HashMurmur2A) == sizeof(HashMurmur2APod) );
BX_FORCE_INLINE void mmix(uint32_t& _h, uint32_t& _k)
{
constexpr uint32_t kMurmurMul = 0x5bd1e995;
constexpr uint32_t kMurmurRightShift = 24;
_k *= kMurmurMul;
_k ^= _k >> kMurmurRightShift;
_k *= kMurmurMul;
_h *= kMurmurMul;
_h ^= _k;
}
static void mixTail(HashMurmur2APod& _self, const uint8_t*& _data, int32_t& _len)
{
while (_len
&& ( (_len<4) || _self.m_count)
)
{
_self.m_tail |= (*_data++) << (_self.m_count * 8);
_self.m_count++;
_len--;
if (_self.m_count == 4)
{
mmix(_self.m_hash, _self.m_tail);
_self.m_tail = 0;
_self.m_count = 0;
}
}
}
static void addAligned(HashMurmur2APod& _self, const void* _data, int32_t _len)
{
const uint8_t* data = (const uint8_t*)_data;
_self.m_size += _len;
mixTail(_self, data, _len);
while(_len >= 4)
{
uint32_t kk = *(uint32_t*)data;
mmix(_self.m_hash, kk);
data += 4;
_len -= 4;
}
mixTail(_self, data, _len);
}
BX_FORCE_INLINE void readUnaligned(const void* _data, uint32_t& _out)
{
const uint8_t* data = (const uint8_t*)_data;
if (BX_ENABLED(BX_CPU_ENDIAN_BIG) )
{
_out = 0
| data[0]<<24
| data[1]<<16
| data[2]<<8
| data[3]
;
}
else
{
_out = 0
| data[0]
| data[1]<<8
| data[2]<<16
| data[3]<<24
;
}
}
static void addUnaligned(HashMurmur2APod& _self, const void* _data, int32_t _len)
{
const uint8_t* data = (const uint8_t*)_data;
_self.m_size += _len;
mixTail(_self, data, _len);
while(_len >= 4)
{
uint32_t kk;
readUnaligned(data, kk);
mmix(_self.m_hash, kk);
data += 4;
_len -= 4;
}
mixTail(_self, data, _len);
}
void HashMurmur2A::add(const void* _data, int32_t _len)
{
HashMurmur2APod& self = *(HashMurmur2APod*)this;
if (BX_UNLIKELY(!isAligned(_data, 4) ) )
{
addUnaligned(self, _data, _len);
return;
}
addAligned(self, _data, _len);
}
uint32_t HashMurmur2A::end()
{
constexpr uint32_t kMurmurMul = 0x5bd1e995;
mmix(m_hash, m_tail);
mmix(m_hash, m_size);
m_hash ^= m_hash >> 13;
m_hash *= kMurmurMul;
m_hash ^= m_hash >> 15;
return m_hash;
}
} // namespace bx

View File

@@ -93,7 +93,7 @@ TEST_CASE("HashAdler32", "")
#define mmix(h,k) { k *= m; k ^= k >> r; k *= m; h *= m; h ^= k; }
uint32_t MurmurHash2A ( const void * key, int len, uint32_t seed )
uint32_t MurmurHash2A(const void * key, int len, uint32_t seed = 0)
{
const uint32_t m = 0x5bd1e995;
const int r = 24;
@@ -148,3 +148,12 @@ TEST_CASE("HashMurmur2A", "")
REQUIRE(test.murmur2a == MurmurHash2A(test.input, bx::strLen(test.input), seed) );
}
}
TEST_CASE("HashMurmur2A-Separate-Add", "")
{
bx::HashMurmur2A hash;
hash.begin();
hash.add("0123456789");
hash.add("abvgd012345");
REQUIRE(MurmurHash2A("0123456789abvgd012345", 21) == hash.end() );
}